Plants can become incorporated into the sediments of virtually any environment, from the oozes of abyssal plains to the silts and sands of delta fronts to brecciated mudflows of volcanic origin. However there is a much narrower range of sedimentary environments in which identifiable plant remains are found in abundance. Generally speaking these are the very shallow or subaerial portions of deltas and estuaries, the channels and floodplains of fluvial systems, lakes of all sizes, ash-falls, and mass-flow deposits such as mudflows. For the purposes of this paper peat swamps are considered as unusual subtypes of deltaic and fluvial environments in which clastic input is low relative to organic accumulation.

The fossil record of plants principally results from individuals, variously sampled from one or more communities, that generally are associated with a depositional site. The processes responsible for the transfer of plant parts from the biosphere to the lithosphere may be biotic (e.g. physiological, animal herbivory, etc.) or abiotic (e.g. traumatic loss by wind and/or water; Krassilov, 1975) in nature. The factors ultimately responsible for the preservation of these plant parts are sedimentological and geochemical. The goals of the subdiscipline of Taphonomy, the study of processes of preservation and how they affect information in the fossil record (Behrensmeyer and Kidwell, 1985), are to understand fossilization through necrology, biostratinomy, and diagenesis (Figure 1). The purpose of this chapter is to provide an overview of taphonomic factors responsible for the generation and preservation of plant macrodetritus. It is not meant to be an exhaustive dissertation or definitive discussion on phytotaphonomy.

Sampling of plant megafossil assemblages has been traditionally a somewhat haphazard process. There has been an almost universal desire to collect the most perfectly preserved, complete, largest (or most conveniently transported) or aesthetically pleasing specimens (I too plead guilty here), and yet collections tend to have been made from the most accessible parts of an outcrop without due regard to depositional geometry, or sedimentological characteristics. Imperfectly preserved, and/or fragmented material typically has been ignored and successive collections frequently have been made from particular beds or horizons that are know to yield “good material” at a particular site. Chance tends to play a significant role in prospecting for new plant beds and heavy reliance is put on the random hit by the hammer (as often as not on float). Once a concentration of plant remains has been found little note is taken of where each recovered specimen was located in relation to other specimens or the entombing sediments.

Statistical methods can assist the paleoecologist in evaluating tentative conclusions and hypotheses that are obtained by more traditional means of analysis. They may also enable the paleoecologist to obtain more information from a set of data than she might otherwise. The purpose of this presentation is to introduce a number of statistical methods that are useful in the analysis of some specific paleoecological problems.

Multivariate analysis can be defined as an analysis in which many variables are examined simultaneously. These types of analyses are invaluable in many scientific fields in which patterns and trends in complex systems must be studied. Multivariate analysis is used in community ecology and paleoecology to summarize patterns of variation between and within different communities so as to allow the ecologist to describe the major patterns and interpret the cause of variation. Some examples of plant paleoecological studies which have used multivariate techniques are Oltz (1969, 1971), Clapham (1971), Robichaux and Taylor (1977), Pheifer (1979), Spicer and Hill (1979), Frederickson (1981), Phillips and DiMichele (1981), Boulter and Hubbard (1982), LaPasha and Miller (1984), Farley and Dilcher (1986), and Kovach (1985, 1987).

Stratigraphic successions provide basic data on past changes in the biosphere. The fossil record may reflect climatic, evolutionary or ecosystem alterations. Careful analysis of Mesozoic and Cenozoic stratigraphic sequences has provided a clearer picture of the diversity of early angiosperms (Doyle, 1978, 1984; Doyle and Hickey, 1976; Crane et al., 1986), their environments of deposition (Hickey and Doyle, 1977), the nature of plant response to Cretaceous-Tertiary boundary events (Wolfe, 1987), the changes in angiosperm fruit and seed sizes as correlated with the increase in forest dominance and changes in dispersal mode over time (Tiffney, 1984, 1986; Wing and Tiffney, 1987), and changes in floristic and physiognomic forest compositions correlated with climatic fluctuations (Leopold and MacGinitie, 1972; Wolfe, 1978, 1981b).

In this discussion, I wish to outline the basic methods of analyzing the paleoenvironmental occurrence of palynomorphs and how this occurrence can be used in paleoecologic interpretation. (Unless otherwise stated, environment as used in this discussion means sedimentary environment.) Palynomorphs have advantages and disadvantages relative to using megafloral remains for the same purpose. Advantages of palynomorphs include their occurrence in large numbers extractable from small volumes of sediment; the usefulness of these large populations for statistical analysis; their occurrence in strata lacking megafossils; and the occurrence of palynomorphs of plants whose preservation potential as megafossils is poor (e.g., fungi, herbaceous plants, algae). Disadvantages of palynomorphs include potential for long-distance transport and reworking; uncertain affinities of many palynomorphs; and bias against zoophilous plants or plants whose pollen contains little sporopollenin (e.g., Populus). Palynomorphs do provide a complementary view of the vegetation which can be of great value if properly used, and whose potential should not be ignored. The best possible analysis would combine megafloral and palynological data at the same level of resolution. Variation of palynology with depositional environment, furthermore, can provide information on individual evolution, community evolution, biostratigraphy, and local variability within depositional environments as well as paleoecology.

Barring the immense and growing literature on Quaternary pollen analysis, the vast majority of the work on the paleoecology of terrestrial systems has concentrated on the Neogene floras of the Pacific Northwest. Study of these floras began with the pioneering exploratory surveys in the last quarter of the 19th century and assumed essentially its modern form with the work of R.W. Chaney, initiated in the 1920, the detailed contributions of H.D. MacGinitie, the continuing productivity of D.I. Axelrod, and, most recently, J. Wolfe. Such a short listing necessarily leaves out the many contributions of patient floristic monographers for I wish to concentrate on floristic treatments that include a major paleoecological element.